DESCRIPTION: HIV establishes latent infection, hampering efforts for a complete cure. The possibility of such a cure rests on new strategies to identify, quantify and ultimately eliminate latently infected cells. An ideal approach would be to identify latently infected cells and eliminae them without the need to reactivate the latent virus. Here, we posit the existence of a novel marker of latently infected cells, based on new insights into HIV and SIV transcription and translation. Specifically, we hypothesize that latent infection is associated with increased levels of viral antisense transcription and translation. Published data on HIV and our preliminary data with SIV infection of rhesus macaques indicate that antisense transcription in infected CD4+ T cells is more prominent when the cells are less activated, and that many of these unique transcripts are polyadenylated and could be translated. Further, we have identified T cell responses mounted during SIV and HIV infection that target surprisingly conserved peptides derived from translation of antisense transcripts. To test our hypothesis, in Aim 1, we will compare viral transcription (sense and antisense) between latently infected CD4 T cells and activated and infected CD4 T cells using both SIV-infected rhesus macaque and HIV-1 infected human cells. To do this, we will use a novel strand-specific deep sequencing approach. Results from Aim 1 will focus our efforts in Aim 2. In Aim 2, we will use T cells that recognize antisense-encoded epitopes and measure their ability to kill latently infected vs. activated and infected CD4 T cells. In sum, this proposal has the potential to identify a novel marker of latently infecte cells and suggest a means by which to eliminate them. These data could prove important to efforts to develop a cure for HIV.